/dports/graphics/vapoursynth-waifu2x-ncnn-vulkan/vapoursynth-waifu2x-ncnn-vulkan-r4/deps/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 128 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 197 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 244 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 313 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 250 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 319 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 350 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 419 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 385 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_neon() local 454 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/graphics/waifu2x-ncnn-vulkan/waifu2x-ncnn-vulkan-20210521/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 128 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 197 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 244 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 313 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 250 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 319 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 350 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 419 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 385 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_neon() local 454 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/benchmarks/vkpeak/vkpeak-20210430/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 128 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 197 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 244 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 313 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 250 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 319 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 350 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 419 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 385 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_neon() local 454 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/misc/ncnn/ncnn-20211208/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 128 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 197 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 244 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 313 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 250 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 319 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 350 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 419 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4to1.h | 385 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_neon() local 454 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_neon() local
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/arm/ |
H A D | convolution_3x3_pack4to1_bf16s.h | 128 … float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4to1_bf16s_neon() local 197 …float32x4_t _tmp34b = vmlaq_n_f32(vmlsq_n_f32(vmulq_n_f32(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack4to1_bf16s_neon() local
|
H A D | convolution_3x3_pack8to1_fp16s.h | 244 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local 313 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to1_fp16sa_neon() local
|
H A D | convolution_3x3_pack8to4_fp16s.h | 250 … float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local 319 …float16x8_t _tmp34b = vfmaq_n_f16(vfmsq_n_f16(vmulq_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.… in conv3x3s1_winograd64_pack8to4_fp16sa_neon() local
|
H A D | convolution_3x3_pack4_fp16s.h | 350 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_r01, 0.5f), _r03, 2.5f), _r05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local 419 … float16x4_t _tmp34b = vfma_n_f16(vfms_n_f16(vmul_n_f16(_tmp01, 0.5f), _tmp03, 2.5f), _tmp05, 2.f); in conv3x3s1_winograd64_pack4_fp16sa_neon() local
|
/dports/graphics/realsr-ncnn-vulkan/realsr-ncnn-vulkan-20210210/src/ncnn/src/layer/x86/ |
H A D | convolution_3x3_pack8.h | 351 …__m256 _tmp34b = _mm256_fmadd_1_ps(_mm256_fmrsub_1_ps(_mm256_mul_ps(_r01, _mm256_set1_ps(0.5f)), _… in conv3x3s1_winograd64_pack8_avx() local 420 …__m256 _tmp34b = _mm256_fmadd_1_ps(_mm256_fmrsub_1_ps(_mm256_mul_ps(_tmp01, _mm256_set1_ps(0.5f)),… in conv3x3s1_winograd64_pack8_avx() local
|